T2 IRC Log: 2010-08-11

This is the log as captured by an IRC bot in the channel. The statements are those of the individual people and might not neccessarily reflect the policy and legal rules as set forth by the T2 SDE Project.

« prev | next »

--- Log opened Wed Aug 11 00:00:49 2010
00:09 < rconstruct> more patches sent....
00:09 < rconstruct> for lprof, gnofract3d, libquicktime, geeqie, inkscape
00:09 < rconstruct> *gnofract4d
00:09 < rconstruct> dvdauthor
00:21 -!- soopurman [~L7-loser@127.sub-75-223-79.myvzw.com] has quit [Ping timeout: 248 seconds]
01:27 < rconstruct> anyone seen this? www.libpng.org/pub/png/libpng.html
01:27 < rconstruct> better to update the png pkg to 1.2.44
01:30 -!- ch172 [~chatzilla@213.160.61.86] has quit [Ping timeout: 258 seconds]
01:37 -!- ch172 [~chatzilla@213.160.61.86] has joined #t2
04:07 -!- old_gaffer [~quassel@211.238.240.87] has quit [Read error: Connection reset by peer]
04:09 -!- old_gaffer [~quassel@211.238.240.87] has joined #t2
05:43 -!- rconstruct [~cgwork@bl17-176-83.dsl.telepac.pt] has quit [Quit: leaving]
10:29 <@rxr> moin
10:36 < koan> rxr: do you have any idea about the mesa error I got yesterday evening?
10:38 <@rxr> did not notice that msg, yet
10:38 <@rxr> hm, guess some glproto dri2proto out of sync or so?
10:39 <@rxr> used to build recently, ..
10:39 <@rxr> but I updated some xorg stuff yesterday
10:39 < koan> this is t2 trunk
10:39 < koan> ok maybe an update will help
10:39 <@rxr> with yesterday's xorg updates ?
10:39 < koan> I updated yesterday afternoon, maybe evening
10:39 <@rxr> maybe you have those updates then
10:40 <@rxr> guess the glproto or so update broke mesa :-(
10:40 <@rxr> will test build on one vm
10:41 < koan> it seems I have built the newest revision (37156), so I had those updates
10:41 <@rxr> you could try svn up some rev before those updates
10:42 <@rxr> and then ./scripts/Create-ErrList -cfg ... -newremove
10:42 <@rxr> should remove the log files of changed packgaes
10:42 <@rxr> so only that handful is rebuild on your nest Build-Target invokation
10:42 < koan> stupid question: how do I update to a specific revision?
10:43 <@rxr> svn up -r ...
10:43 <@rxr> can even use dates svn up -r {2010-08-01}
10:43 < koan> thanks, I'm choosing 37139, right before the xorg updates
10:43 <@rxr> I think dates are a litte fuzzy, timezones and start vs. end of day :-)
10:45 < koan> ok, I'll see what the build gives now
10:46 <@rxr> do you build i386 or x86-64 ?
10:48 < koan> x86-atom
10:49 <@rxr> scripts/Cleanup -build build/minimal-xorg-9.0-trunk-generic-x86-64-linux/
10:50 < koan> the mesa build is busy now, holding my breath :-)
10:50 <@rxr> for netbook, atom is it probably favourable to use size optimizations:
10:50 <@rxr> (size) Hard optimise for size (sometimes faster)
10:51 < koan> why?
10:51 <@rxr> more compact code, less memory, cache trashing, etc.
10:51 <@rxr> I even use to build my server targets Os, ...
10:51 <@rxr> I was thinking about making it the default, ...
10:52 < koan> and this is a significant difference?
10:52 <@rxr> Os is like O2, just without the "excessively blowing up code" options
10:52 <@rxr> -Os enables all -O2 optimizations that do not typically increase code size. ...... It may, however, yield faster code for programs that do not require the ...
10:52 <@rxr> ^- put that in quotes
10:54 < koan> hm still the same build faliure
10:54 < koan> should I cleanup the build first?
10:54 <@rxr> ok, guess it's time to look at that codeline and see what is being used
10:55 <@rxr> no newremove is enough, I guess it build the relevant packages after it?
10:55 < koan> yes
10:55 <@rxr> you can open the erroring file in an editor and see what's going on at the line
10:55 <@rxr> my build should be there in some hour, too
10:55 <@rxr> Error logs from minimal-xorg-9.0-trunk-generic-x86-atom-linux:
10:55 <@rxr> 473 builds total, 6 completed fine, 0 with errors, 467 to go.
10:58 < koan> hm I'll take a look at it later, first some work to do (building T2 in the background is not good for my writing efficiency :-))
10:59 <@rxr> I can imagine
10:59 <@rxr> .oO(many patches on the list)
10:59 <@rxr> guess it's merge time
11:01 < koan> btw, I booted a minimal T2 on my Acer Aspire One yesterday, and it really boots fast :-) I'll create a minimal desktop target for my netbook with awesome, wicd, and so on, this will be much faster than the Ubuntu Netbook Bloat I used
11:02 <@rxr> yeah, guess we really should bring out some "userfriendly T2 remix"
11:02 <@rxr> http://www.userfriendly.org/
11:02 <@rxr> like graphical install and such
11:03 < koan> I can contribute the netbook target when I have it finished and tested, although it will be a minimalistic netbook distribution, I'm not interested in a full desktop environment, especially on a netbook
11:03 <@rxr> guess most people will also not fall in love with a tiling wm ..
11:04 < koan> maybe openbox
11:06 <@rxr> ah, finally found out why the nightly t2 update tracker produced so much already udated packages, was still running on the 8.0 tree.... :-)
11:27 < CIA-32> rene * r37157 /trunk/package/lua/lua-posix/lua-posix.desc: * updated lua-posix (5.1.4 -> 5.1.6)
11:39 < CIA-32> rene * r37158 /trunk/package/network/netkit-rsh/ (netkit-rsh.conf rlogind.patch):
11:39 < CIA-32> luisbf@gmx.com:
11:39 < CIA-32> * fixed netkit-rsh for builds without pam and gcc >= 4.3
11:41 < CIA-32> rene * r37159 /trunk/package/graphic/scribus/scribus.desc:
11:41 < CIA-32> luisbf@gmx.com:
11:41 < CIA-32> * removed latest scribus Download tag workaround, apparently the
11:41 < CIA-32> filename was fixed on the upstream download site
11:42 < CIA-32> rene * r37160 /trunk/package/multimedia/live/live.desc:
11:42 < CIA-32> luisbf@gmx.com:
11:42 < CIA-32> * updated live (2010.07.13 -> 2010.07.29)
11:47 < CIA-32> rene * r37161 /trunk/package/multimedia/libmikmod/libmikmod.desc:
11:47 < CIA-32> luisbf@gmx.com:
11:47 < CIA-32> * updated libmikmod (3.1.10 -> 3.1.11)
11:52 < CIA-32> rene * r37162 /trunk/package/multimedia/libmikmod/libmikmod.desc: * mark libmikmod AUTOMAKE-QUIRK to build on x86-64, et al.
11:53 < CIA-32> rene * r37163 /trunk/package/printing/foomatic-filters/foomatic-filters.desc: * updated foomatic-filters (3.0-20081204 -> 4.0.5)
12:56 < CIA-32> rene * r37164 /trunk/package/multimedia/libcss/libcss.desc: * marked libcss AUTOMAKE-QUIRK for x86-64, et al.
13:05 < CIA-32> rene * r37165 /trunk/package/multimedia/xvidcap/ (shmproto_inclusion.patch xvidcap.cache):
13:05 < CIA-32> luisbf :
13:05 < CIA-32> * fixed xvidcap to build with altest xorg headers
13:12 < CIA-32> rene * r37166 /trunk/package/graphic/libxmi/libxmi.desc: * marked libxmi AUTOMAKE-QUIRK, for x86-64, et al.
13:12 <@rxr> http://www.t2-project.org/packages/libxmi.html
13:19 < CIA-32> rene * r37167 /trunk/package/graphic/libopenraw/libopenraw.desc:
13:19 < CIA-32> luisbf :
13:19 < CIA-32> * updated libopenraw (0.0.5 -> 0.0.8)
13:27 < CIA-32> rene * r37168 /trunk/package/gnome2/gtkimageview/ (. gtkimageview.cache gtkimageview.desc):
13:27 < CIA-32> luisbf@gmx.com:
13:27 < CIA-32> * added gtkimageview (1.6.4) - A simple image viewer widget for GTK+
13:27 < CIA-32> rene * r37169 /trunk/package/graphic/ufraw/ufraw.desc:
13:27 < CIA-32> luisbf :
13:27 < CIA-32> * updated ufraw (0.15 -> 0.17)
13:35 < CIA-32> rene * r37170 /trunk/package/graphic/gnofract4d/gnofract4d.desc:
13:35 < CIA-32> luisbf :
13:35 < CIA-32> * updated gnofract4d (3.12 -> 3.13)
13:35 < CIA-32> rene * r37171 /trunk/package/multimedia/dvdauthor/dvdauthor.desc:
13:35 < CIA-32> luisbf :
13:35 < CIA-32> * updated dvdauthor (0.6.14 -> 0.6.18)
13:36 < CIA-32> rene * r37172 /trunk/package/multimedia/libquicktime/libquicktime.desc:
13:36 < CIA-32> luisbf :
13:36 < CIA-32> * updated libquicktime (1.1.3 -> 1.1.5)
13:37 <@rxr> http://t2-project.org/packages/vigra.html
13:50 < CIA-32> rene * r37173 /trunk/package/graphic/lprof/bool.patch:
13:50 < CIA-32> luisbf :
13:50 < CIA-32> * fixed lprof to compile due missing BOOL
13:52 < CIA-32> rene * r37174 /trunk/package/graphic/geeqie/ (. geeqie.cache geeqie.desc):
13:52 < CIA-32> luisbf :
13:52 < CIA-32> * added geeqie (1.0) - A graphics file viewer, forked from GQView
13:58 <@rxr> http://www.t2-project.org/packages/geeqie.html
14:23 <@rxr> Error logs from minimal-xorg-9.0-trunk-generic-x86-atom-linux:
14:23 <@rxr> 473 builds total, 139 completed fine, 0 with errors, 334 to go.
14:25 < koan> how do you get these statistics?
14:25 <@rxr> scripts/Create-ErrList -cfg ...
14:25 <@rxr> VM feels slow today
14:25 < koan> I can run this safely while the build is running?
14:25 <@rxr> getting the status: yes
14:26 <@rxr> with the -newdelete etc. it can by times interfere the currently building package (if files get removed from an changed package) but I usually just do it with much interferen e
14:27 <@rxr> I was also thinking about modinfing Build-Target (and maybe Emerge) to print how much percent of the packages are build at the "finished line" of each package
14:27 < koan> that would be interesting, yes, especially for Build-Target
14:59 < CIA-32> rene * r37175 /trunk/package/gnome2/gtkglext/deprecatedmacros.patch:
14:59 < CIA-32> luisbf :
14:59 < CIA-32> * fixed gtkglext not to use deprecated macros
15:01 <@rxr> the new (not yet comitted) make is really picky:
15:01 <@rxr> == 14:58:53 =[5]=> Building gnome2/gstreamer [0.10.30 9.0-trunk].
15:01 <@rxr> !> Makefile:939: *** missing separator (did you mean TAB instead of 8 spaces?). Stop.
16:23 < CIA-32> rene * r37176 /trunk/package/network/netperf/ (netperf.conf netperf.desc): * updated netperf (2.4.1 -> 2.4.5)
16:56 <@rxr> ouhm
16:56 <@rxr> looks like virtio net performance is pretty bad, damnit
16:57 <@rxr> Atom 330 server to 6 core Phenom X2:
16:57 <@rxr> Size Size Size Time Throughput
16:57 <@rxr> bytes bytes bytes secs. 10^6bits/sec
16:57 <@rxr> 87380 16384 16384 10.00 829.34
16:58 <@rxr> 8 core xeon:
16:58 <@rxr> 87380 16384 16384 10.45 18.84
16:58 <@rxr> ok, this older xeon has absymal VT performance anyway, but a decent Phenom VM does only look slightly better :-(
17:07 <@rxr> native performance of that 8-way xeon:
17:07 <@rxr> 87380 16384 16384 10.02 940.78
17:12 -!- rconstruct [~cgwork@bl15-63-139.dsl.telepac.pt] has joined #t2
17:12 < rconstruct> hi
17:13 < rconstruct> doing a build from scratch with latest svn
17:13 < rconstruct> might have some more patches
17:13 <@rxr> hi rconstruct
17:13 < rconstruct> hi rxr :)
17:13 <@rxr> I think I applied some 99% of yours :-)
17:13 < rconstruct> thanks for the suggestion, wasn't aware of the AUTOMAKE-QUIRK
17:14 < rconstruct> just checked it, doing a build atm :)
17:14 < rconstruct> actually, have one completed yesterday
17:14 <@rxr> :-)
17:14 < rconstruct> but doing another one just to recheck everything and test a couple of other things
17:14 <@rxr> btw. do you exclusively use ROCK, now T2, or do you have other OSs running beside T2?
17:15 < rconstruct> rock 2.03 in a old athlonxp
17:15 < rconstruct> t2 in everything else
17:15 < rconstruct> couple of athlonxps, athlon64x2, and phenomII x6
17:15 <@rxr> nice to hear
17:15 < rconstruct> only exception is laptop with ubuntu
17:15 <@rxr> our router is still - # cat /etc/ROCK-VERSION
17:15 <@rxr> ROCK Linux 2.0.1 (2004/03/13)
17:16 < rconstruct> ehehehe
17:16 <@rxr> (with security updates where reasonable)
17:16 < rconstruct> speaking about security, that libpng issue was creepy
17:17 < rconstruct> don't know which was the first rock version i used, but i was using redhat 6.2 iirc
17:17 <@rxr> native phenom II X4 to X6 performance, only onle local switch:
17:17 <@rxr> 87380 16384 16384 10.02 941.24
17:18 < rconstruct> any way to get full build stats?
17:18 < rconstruct> after a complete scratch build test?
17:18 <@rxr> what kind of stats?
17:19 <@rxr> scripts/Create-ErrList not enough?
17:20 < rconstruct> pkg time, full time, cache used, etc... but nvm
17:20 <@rxr> nop
17:21 <@rxr> you could hack one, htough
17:21 <@rxr> though
17:22 < rconstruct> lol
17:22 < rconstruct> btw, the inkscape patch, they should be releasing 0.48 soon, but in any case, things should build, hopefully 0.48 will deal gracefully with gcc 4.5
17:22 < rconstruct> ok, going to start a build, brb, need to check config
17:24 < rconstruct> the only thing that worries me is autoconf 2.66/2.67
17:26 <@rxr> yeah, auto* junk is always to be worried about, sigh
17:26 <@rxr> cu
17:28 < rconstruct> i'll b around, will have more patches though
17:28 < rconstruct> see you then
17:32 <@rxr> ok, network performance on an phenom II X4 is significanly better than on that xeon:
17:32 <@rxr> 87380 16384 16384 10.00 366.48
17:32 <@rxr> not shining, but better, ...
17:33 < rconstruct> btw, you can build poppler, with autoconf 2.67
17:33 < rconstruct> and pkgconfig without the pkg.m4 patch to fix FDdnl?
17:37 <@rxr> I think so
17:37 <@rxr> interesting, with hpet clocksource the performance is a little better?
17:37 <@rxr> 87380 16384 16384 10.03 425.16
17:39 < rconstruct> overclocked this a bit
17:39 < rconstruct> @ 3.4gHz
17:39 < rconstruct> to get all 4 dimms at 1600mHz
17:40 < rconstruct> i'm not a fan of overclocking, i usually underclock things, not OC them
17:40 < rconstruct> just stress-testing the machine
17:40 < rconstruct> if it's going to blow, then at least it'll be during the 30 days "grace period"
17:40 < rconstruct> for delivery/replacement
17:45 < rconstruct> hmm, trouble
17:45 < rconstruct> now audiofile.m4
17:46 < rconstruct> i'm thinking Aldas might be right about downgrading to 2.65
17:47 < rconstruct> well, FDdnl can be patched in audiofile.m4 too
17:47 < rconstruct> let's hope there aren't many m4 files to fix lol
17:50 <@rxr> yeah, coordinate with aldas what works best :-)
17:54 <@rxr> aha:
17:54 <@rxr> 87380 16384 16384 10.01 542.65
17:54 <@rxr> hpet clocks all over
18:07 < rconstruct> well, autoconf 2.66/2.67 really mean trouble with some pkgs
18:08 < rconstruct> not much choices here, either downgrading to 2.65, or keeping 2.67 but patching some m4 files as trouble appears
18:17 < rconstruct> k, going to do a build with 2.65
18:54 -!- mjungwirth2 [~mjungwirt@213.174.252.112] has joined #t2
18:56 < rconstruct> rxr: patches sent
19:04 <@rxr> nice
19:04 <@rxr> guess I have to start reviewing t2 patches at home at morning and night
19:04 <@rxr> otherwise I do not archive much else during office hours :-)
19:07 < rogermason> Oops, sorry, I just sent in two more.
19:07 <@rxr> that's fine, night is long
19:08 < rogermason> Well, one of them is so slow to build, you could at least have a nap while waiting.
19:10 <@rxr> well, I got some free cores, here
19:21 < CIA-32> rene * r37177 /trunk/package/scientific/xmds/ (. xmds.cache xmds.desc):
19:21 < CIA-32> Roger Mason :
19:21 < CIA-32> * added xmds (1.6.6) - eXtensible Multi-Dimensional Simulator
19:26 < rconstruct> crap, forgot one
19:26 < rconstruct> lol
19:26 < rconstruct> djvulibre
19:26 < rconstruct> well, two, gv
19:27 <@rxr> hm ?
19:27 <@rxr> ah
19:27 < rconstruct> patches
19:27 < rconstruct> but it's better to wait a few hours
19:27 < rconstruct> and i'll send in packs of 12
19:27 < rconstruct> lol
19:27 <@rxr> no just send now
19:27 <@rxr> better to keep track if there are few patches per email
19:27 < rconstruct> ok :)
19:27 <@rxr> especially if I do not apply one instantly and keep thinking about it
19:28 < rconstruct> the pfstools/pfstmo/pfscalibration, been using & updating that in my custom tree for some years, never had any issues on x86 nor x86_64, but haven't tried in other archs
19:33 < rconstruct> k, posted
19:38 <@rxr> 0 9.0-trunk].
19:38 <@rxr> -> Parallel build using 7 jobs enabled.
19:38 <@rxr> -> JAVA_HOME == 08/11/10 19:38:09 =[5]=> Finished building package paraview.
19:38 <@rxr> err
19:38 <@rxr> == 19:15:15 =[5]=> Building scientific/paraview [3.8.0 9.0-trunk].
19:38 <@rxr> == 08/11/10 19:38:09 =[5]=> Finished building package paraview.
19:38 <@rxr> that is a 8 core XEON, however that VM only got 7
19:38 <@rxr> and it's virtualization performance sucks due to lack of nested paging, ...
19:39 <@rxr> guess my newer, 6-core AMD VM woul build it quicker
19:39 <@rxr> at least I hope so, should probably give it a try for the benchmarking of it :-)
19:40 < CIA-32> rene * r37178 /trunk/package/scientific/paraview/ (. paraview.cache paraview.conf paraview.desc):
19:40 < CIA-32> Roger Mason :
19:40 < CIA-32> * added paraview (3.8.0) - Parallel Visualization Application
19:41 < rogermason> was that // build using distcc?
19:41 <@rxr> no distcc
19:41 <@rxr> native cores
19:45 < rogermason> OK. On my first attempt (using distcc) it seemed to get locked in an unending cycle during the linking phase. The distcc host doing most of the work became uncontactable at some point in the build. Dunno if that is a distcc, paraview or other problem.
19:46 <@rxr> I think linking is not distributed via distcc anyway
19:47 <@rxr> only object compoilation
19:47 <@rxr> that is gcc invocation, not ld
19:47 <@rxr> because it had to send all the objects and linking is mostly soldering binaries together instead of number crunching as compiling ist
19:49 < CIA-32> rene * r37179 /trunk/package/graphic/djvulibre/djvulibre.desc:
19:49 < CIA-32> luisbf :
19:49 < CIA-32> * updated djvulibre (3.5.20 -> 3.5.22)
19:49 < CIA-32> rene * r37180 /trunk/package/x11/gv/gv.desc: * updated gv (3.6.3 -> 3.6.7)
19:50 <@rxr> rconstruct: is that gv update stable?
19:50 <@rxr> the 7 makes me nervous
19:50 < rconstruct> lol
19:51 < rconstruct> well, it's trunk/9.0, but i will find out in a few hours
19:51 < rconstruct> checking Aldas suggestion to use 2.66 + hotfix
19:51 < rconstruct> autoconf
19:51 < rconstruct> still have the same problem with poppler
19:51 < rconstruct> possibly undefined macro: AS_MESSAGE_LOG_FDdnl
19:51 < rogermason> rxr: Yes, that is my understanding too. I'm really not sure what went wrong.
19:53 < CIA-32> rene * r37181 /trunk/package/x11/gv/gv.desc: * updated gv (3.6.7 -> 3.6.9)
19:54 < rconstruct> 3.6.9? how did i miss that
19:54 <@rxr> last even numbered one until I know if 3.7 is considered stable :-)
19:54 < rconstruct> ah k
19:54 < rconstruct> btw, updated audiofile
19:57 < CIA-32> rene * r37182 /trunk/package/graphic/libpng/libpng.desc:
19:57 < CIA-32> luisbf :
19:57 < CIA-32> * updated libpng (1.2.42 -> 1.2.44) - SECURITY
19:57 <@rxr> you are good in updating, aren't you?
19:57 < rconstruct> lol
19:57 < rconstruct> well, might as well try and see how things behave
19:57 <@rxr> want an svn write access acount for it? so I save cut'n pasting all the patches ?
19:57 < rconstruct> lol, perhaps later, we can talk about it later :)
19:58 < rconstruct> ok, i ran into the same issue with 2.66+hotfix, 2.66, 2.67, autoconf
19:58 < rconstruct> going to build from 0 with autoconf 2.65
19:58 <@rxr> will merge the other things later
19:58 <@rxr> gotta go - cu then
19:58 < rconstruct> sure, see you later then
22:43 -!- mjungwirth2 [~mjungwirt@213.174.252.112] has quit [Quit: Nettalk6 - www.ntalk.de]
--- Log closed Thu Aug 12 00:00:49 2010